A Terrain Relative Navigation Sensor Enabled by Multi-Core Processing

Terrain Relative Navigation (TRN) provides accurate position estimates to spacecraft for precision planetary landing and autonomous primitive body exploration. A bolt-on instrument that provides the sensing and computing required for TRN will result in more accurate and robust position estimates and will simplify TRN validation. Multi-core processors provide the significant computational capability required for TRN, are straightforward to program and are being developed for space applications. We have implemented two versions of TRN on a multi-core processor and tested them in a laboratory setting. For primitive-body navigation we have demonstrated 4 second TRN updates with accuracies on order 1% of altitude. For a Mars landing application we have shown two second updates while taking out kilometer scale position uncertainties.